February 2008 - Inkscape Devel

feDisplacementmap fix
by jf barraud 07 Feb '08

07 Feb '08

Hi all! Atm, feDisplacementmap output depends on zoom factor (see here<https://bugs.launchpad.net/inkscape/+bug/182059> ). I had a look at the code to try to fix this, but I need someone to validate the code. Precise meaning of area, unit transforms, and particulraily of the function "enlarge_area" are not perfectly clear to me. The modifs are all in nr-filter-displacement-map.cpp.<http://launchpadlibrarian.net/11804938/nr-filter-displacement-map.cpp> I just changed "scale" with "scalex" and "scaley" defined as: Matrix trans = units.get_matrix_primitiveunits2pb(); double scalex = scale*trans.expansionX(); double scaley = scale*trans.expansionY(); where apparently appropriate (in the render() and enlarge_area() methods), but I would feel more confortable if someone having a better understanding of the code could check if this is correct or if other changes are needed elsewhere... Thanks! jfb.

3 2

Re: [Inkscape-devel] 'feComponentTransfer'
by Felipe Sanches 07 Feb '08

07 Feb '08

I am stuck on this. Can anybody help please? I have updated the patch to current svn (r17273) On Jan 17, 2008 3:55 AM, Felipe Sanches <felipe.sanches@...400...> wrote: > I started coding it. Though it is not working yet. > > How does inkscape knows that my Class handles a specific tag? I mean... > how will i have 4 different tags (<feFuncR>, <feFuncG>, <feFuncB> and > <feFuncA> be handled by my SPFeFuncNode class ?) > > I supose that you are more confortable with this than me, so, if > possible, could you take my patch and complete it? I have spent hours trying > to figure out how to do it and now I wish to focus on implementing the > render method. > > > I would also ask for nicholas to help us to adapt the UI > > Hoppefully we can have this filter implemented before 0.46 release. Then > the only ones that would be out are feImage and feTile > > > On Jan 16, 2008 5:57 PM, Niko Kiirala <niko@...1267...> wrote: > > > Wed, 16 Jan 2008 12:45:15 -0200 > > "Felipe Sanches" <felipe.sanches@...400...> kirjoitti: > > > > > can you help me? I am not sure how do we deal with nested tags . How > > > do I read their attributes? > > > > I belive, the best way to set this up is to create SP-classes for those > > tags, that can be children. Then implement child_added, remove_child and > > order_changed callbacks in the parent class. These callbacks are not > > strictly required, but following modifications is often useful. > > > > With these, in the parent class you can access this_object->children, > > this_object->children->next etc. (this_object here is a pointer to > > object of the parent class type) Then, you may need suitable accessors > > in the cild classes, that give you the information you need. > > > > See sp-fespecularlighting.cpp for an example. The child classes in this > > case are in sp-fespotlight.cpp, sp-fedistantlight.cpp and > > sp-fepointlight.cpp > > > > -- > > Niko Kiirala > > niko@...1267... > > > >

1 0

improving a bit more the feImage filter
by Felipe Sanches 07 Feb '08

07 Feb '08

this is my updated patch for hadling svgelements referenced by id on the feImage filter primitive. It is not working yet but is much better than the previous patch I sent to the list. Now the URIReference is properly handled and a NRArena/NRArenaItem created. Then I create a pibblock and render the svg element in it. When I call the nr_arena_item_invoke_render method I get this error: ** (inkscape:6874): WARNING **: rendering element File display/nr-arena-item.cpp line 308 (?): Assertion item->state & NR_ARENA_ITEM_STATE_BBOX failed I dont know how to get rid of this one. I am going to travel now and will be back only next week. If anybody wants to give it a try, this patch needs a last touch to get working. (it also adds an extra button to the feImage settings UI: to use an element instead of an image, you must select the desired element and then click on the "SVG Element" button) best wishes, Juca

1 1

Expected LPEs for 0.47
by unknown＠example.com 07 Feb '08

07 Feb '08

Hi all, JF sent me this and I couldn't resist posting it here: http://math.univ-lille1.fr/~barraud/Inkscape/effectPreview.svg I especially like the sketch effect! Cheers, Johan

7 8

pattern-along-path LPE recent additions
by jf barraud 07 Feb '08

07 Feb '08

Hi, Hum, Johan is right. I did not consider this as a new feature, but I was wrong. Sorry--- still learning! :-( I think it is definitely worth thinking about it: the point is that pattern-along-path loose a lot of usability if the offset parameters are not there (at least the 'space' added between copies): several people asked me about the old python script, just to have this functionality (that is what made me think of this as a kind of bug...). For instance, in many (most?) cases (foot prints, stars, any shape...), you don't want the copies of the pattern to touch each other. Moreover, the work arround are not easy... Making the dimensions proportional or not to the size of the object is codewise straightfoward, but Johan is right stressing that the right decisions have to be made here. This requires some time to think about it, and might be worth postponing to 0.47. On the other hand, it would be a pitty to release lpe pattern-along-path without it... I am just not experienced at all in such big projects as inkscape, and I'll devoutly follow guru's advices! In particular, I'll perfectly agree if we decide to remove this. JF.

1 0

Fillet Tool Blueprint Discussion
by Joel Holdsworth 06 Feb '08

06 Feb '08

Hi All! I'm in the process of getting together a blueprint for a fillet tool that I'm drafting: http://wiki.inkscape.org/wiki/index.php/SpecFilletChamfer. I'd love to get some discussion and feedback going. I've got a couple of questions that I'd like to discuss at the bottom, and I'd like to hear people's feedback. Thanks Joel

9 22

Filter effects UI
by Niko Kiirala 06 Feb '08

06 Feb '08

Hi There are a couple of issues with filter effects UI, that I'm concerned about in the upcoming release. First, at some point of development, the blur slider was disabled when there was a more complex filter attached to the object. I believe that this is a good thing, because it's not too clear, what it should do. Nevertheless, right now it causes a crash. I haven't dwelved much into this problem and there may be some bigger underlying problem, that needs fixing. But unless there's a plan on how this slider should work with complex filter already attached to the object, I believe it should be disabled in those cases. (see https://bugs.launchpad.net/inkscape/+bug/186281 ) Second is a problem I fully realized only today: there's no user interface for setting enable-background property. Without this, the only working input images are Source Graphic and Source Alpha. Background Image and Background Alpha only contain data, when there's property 'enable-background:new' set on some parent object of the filtered object. Without this, they will just appear to not work. (see https://bugs.launchpad.net/inkscape/+bug/189341 ) (Fill Paint and Stroke Paint don't work because there's no renderer support yet for them in Inkscape) -- Niko Kiirala niko@...1267...

3 3

pattern-along-path LPE recent additions
by unknown＠example.com 06 Feb '08

06 Feb '08

Hi all, Recently, JF added support of some more of path-along-path extension's parameter in the pattern-along-path LPE. (spacing, and offset parameters) I am not too certain that this change is good so close to release. Especially I am concerned with ambiguity or "wrong" calculation of parameter values. Could someone please have a (somewhat thorough!) look at what the new parameter values mean and that the resulting path is consistent with that meaning? What happens for example, when scaling the original path? I really think the meaning of their values should depend on the "scale width in units of length" checkbox. It is easy to revert (the svn revision changed only one file and reverting it does not break compilation), so we can postpone the decision _a bit_. Thanks a lot, Johan

1 0

Linux dev. binaries?
by Terry Brown 06 Feb '08

06 Feb '08

Is there a source of current SVN binaries for linux of which I'm not aware? Seems to me a large base of possible testers are being lost by not having these available - I know everyone's hard at work on good stuff, but is there anyone who could re-start the autopackage builds so linux inkscape users can test and report bugs? I just grabbed my early Jan. build to do some work, ran into a real doozy of a bug... couldn't find it on launchpad... figure it must be fixed in SVN... took me quite some time to make an SVN compile (most people wouldn't bother), and of course the bug's gone, great. Had I not spent all that time on an SVN compile my choices would have been either report a bug that's gone, or not report anything at all. Cheers -Terry

1 0

Re: [Inkscape-devel] [Inkscape-user] fonts disappear in OSX
by jiho 06 Feb '08

06 Feb '08

On 2008-February-05 , at 21:40 , Gail Carmichael wrote: > jiho wrote: >> On 2008-February-05 , at 19:12 , churro wrote: >>>> - select the faulty text, open the xml editor and check what the >>>> font >>>> attributes are (name, weight etc.) >>> these are the attributes: I'm using Culrz MT >>> font-size:98.06591034px;font-style:normal;font- >>> weight:normal;fill:#ff8080;fill-opacity:1;stroke:none;stroke-width: >>> 1px;stroke-linecap:butt;stroke-linejoin:miter;stroke-opacity: >>> 1;font-family:Curlz >>> MT >> >> This is a (Microsoft I think) Font Suitcase format, which indeed >> causes issues. The strange thing is that it worked in an earlier >> version of Inkscape. I am guessing that the difference actually >> comes from the font itself. When did you start this piece of work? >> Did you upgrade your system or MS Office in between? > > I've never even heard of these suitcase things (I assume you figured > out the format from the file itself and not the above properties). > Hopefully this means that the code I did has not adversely affected > the situation. I wrote it so that it looks for my new information > first, then falls back onto the old tried tested and true way. OK, I did a little research and there's no wonder you never heard of it: it is a mac specific thing. To get a brief overview of font management on OS X (which I've always thought to be very well designed) you can skim over this: http://docs.info.apple.com/article.html?artnum=25251 http://developer.apple.com/technotes/tn/tn2024.html The Font Suitcase thing is a format dating back from OS 9 (the original Macintosh system, before it was unixified). This is a weird format, in which the whole font is stored in the resource fork and nothing is stored in the data fork. As a result the file appears to be 0 bytes: $ ll Curlz\ MT -rwxrwxr-x@ 1 root admin 0 Apr 7 2004 Curlz MT nor do they have a file extension. The font information however is just regular true type. OS X (i.e. the unixified version) does not ship with these fonts now (I'm not even sure it ever shipped with some) but supports them for backward- compatibility's sake. The only app I know that ships with such fonts is.... you guessed it: MICROSOFT Office. Well you can't ask them to be good players on a platform which is competing with them, for sure. I don't expect support to appear in Pango anytime soon since the format is deprecated. Anyway, it seems perfectly safe to convert those to regular TTF, no information is lost: it just moves it from the resource fork to the data fork. It just means each user has to do it. PS: I would be curious to know which format Office 2008 uses for its font on OS X. Does someone have info on this? PPS: you may see some work on font suitcases in Pango, but this is not (AFAIK) related to this particular format. It is probably related to the new font suitcase format introduced with OS X, which has the dfont extension and is kind of the contrary: everything is in the data fork and the resource fork is empty of font information. This makes working with this format easier. But it is at the same time quite new so support is not yet perfect. These font cause the typeface bugs on Inkscape/OS X, in which some font faces are available in the menu (bold, semi bold etc) but selecting them does not do anything. >> [...] >> >> Please send this the needed info to Gail too, she's the font expert. > > > Unfortunately, I'm not familiar with this suitcase thing you > mentioned. Normally I would be able to look it up and learn about > it etc, but unfortunately I have a heavy course-load for the > semester and probably don't have the necessary time to devote to > it. I am cc'ing the devel list in case anyone else has the > knowledge to have a look. Still, we should try to determine for > sure if it could be an issue with my code, and I can (hopefully) fix > that if it is. Well something changed recently apparently. I hope that with the help of Nick you'll be able to reduce the time frame of occurrence of the bug. Hope that helps. JiHO --- http://jo.irisson.free.fr/

2 4